The NVIDIA Holoscan platform is expected to accelerate the development of software workflows in the medical sector by enabling the development and deployment of AI-powered applications using existing MATLAB algorithm codebases.

MathWorks, a leading developer of mathematical computing software, today announced integration of MATLAB with NVIDIA Holoscan, a real-time AI computing software platform for building medical devices.

This capability enables medical device engineers to analyze streaming data and accelerate the development and deployment of visualization applications by wrapping existing MATLAB algorithms and functions with the GPU-accelerated NVIDIA Holoscan Operator for real-time data processing and inference.

Medical device engineers must keep pace with the rapid pace of innovation in cutting-edge materials and electronics technologies, as well as comply with complex and changing international regulations. This has resulted in many devices becoming obsolete shortly after hitting the market, giving rise to Software as Medical Device (SaMD).

'Software as a Medical Device (SaMD)' refers to a medical device that is not dependent on hardware, has functions that match the intended use of the medical device, and is comprised solely of independent software.

Engineers must develop software-defined workflows that enable additional software features to be integrated after initial deployment to ensure that software medical devices remain competitive in the marketplace.

“The medtech industry is being transformed by artificial intelligence,” said David Niewolny, business development manager for medical technologies at NVIDIA. “NVIDIA and MathWorks are accelerating AI-driven innovation in medtech by providing a highly compatible development environment for the growing MATLAB development community within the medical-grade NVIDIA Holoscan platform.”

NVIDIA Holoscan is a sensor processing platform that simplifies the development and deployment of AI and high-performance computing applications that deliver real-time insights.

It also provides the full-stack infrastructure required for scalable software-defined processing of streaming data at the edge, enabling the introduction of modern AI applications into clinical settings.

Medical device engineers can use existing built-in matrix operations and complex toolbox functions for image and signal processing, filtering, transformations, and deep learning algorithms through the integration of Holoscan and MATLAB.

To implement the Holoscan pipeline in MATLAB, create a MATLAB function and use GPU Coder. The four-step process involves generating accelerated CUDA code, creating a Holoscan Operator wrapper, and rebuilding the Holoscan application using the new MATLAB Operator. The resulting software-defined workflow is fully functional for medical devices, and additional integrated validation and verification capabilities in MATLAB and Holoscan help ensure compliance with industry regulations and standards such as IEC 62304.

“Engineers can now write MATLAB functions with NVIDIA Holoscan and have them run thousands of times faster,” said David Rich, director of product marketing at MathWorks. “With millions of customers looking to design, develop, and test products that comply with industry regulations and standards, our collaboration with NVIDIA, an industry leader, will help drive innovation in medical devices.”